01
Multi-Model Selection
Ten published models, from Hailuo 2.3 at 36 credits a clip to Veo 3.1 at 300. Each is priced per second of output, and the picker shows the number before you spend it.
작동 방식
핵심 기능
02
Cinematic Quality Output
Generate up to 1080p across most of the roster, and 4K on Kling O3, Kling V3.0 and Veo 3.1. Resolution is a per-model cap rather than a plan tier — the table below lists each one.
03
Fine-Grained Controls
Adjust duration, aspect ratio, motion intensity, and camera angles. Use negative prompts to exclude unwanted elements and refine your output.
Every text-to-video model, and what a clip costs
Ten models are published, and the price gap between the ends of the roster is roughly eight-fold. Credits are a flat $0.01 each and video bills per second, so the figure below is the real cost of one five-second clip.
| Model | Price / 5s | Clip length | Max resolution | Holds a cast? |
|---|---|---|---|---|
| Hailuo 2.3 | 36 cr ($0.36) | 6–10s | No | |
| MiniMax H3 | 45 cr ($0.45) | 5–15s | 768P | Yes — 9 refs |
| PixVerse C1 | 49 cr ($0.49) | 1–15s | 1080P | Yes — 7 refs |
| Grok Imagine Video | 53 cr ($0.53) | 4–10s | 720P | Yes — 7 refs |
| Wan 2.6 | 75 cr ($0.75) | 5–15s | 1080P | No |
| Kling O3 | 84 cr ($0.84) | 3–15s | Yes — 4 refs | |
| Kling V3.0 | 95 cr ($0.95) | 3–15s | Yes — 4 refs | |
| HappyHorse 1.0 | 105 cr ($1.05) | 3–15s | 1080P | No |
| Vidu Q3 | 116 cr ($1.16) | 1–16s | 1080P | Yes — 4 refs |
| Seedance 2.0 | 228 cr ($2.28) | 4–15s | 1080P | Yes — 9 refs |
| Veo 3.1 | 300 cr ($3.00) | 8–8s | 4K | Yes — 3 refs |
| Seedance 2.5 | 355 cr ($3.55) | 4–30s | 720P | Yes — 30 refs |
Read from the live model registry, so it moves when prices move. Longer clips scale linearly from the five-second rate.
Using it as a free AI short video generator from text
Signing up grants 100 credits, once, with no card. On Hailuo 2.3 at 36 credits for five seconds that is two clips — enough to judge whether the output is worth paying for, and deliberately not enough to finish a project on.
Reelfo does not stamp a watermark on anything it renders. The compose step copies the video stream through untouched — there is no overlay filter in the pipeline to apply one.
The "holds a cast?" column is the one that matters if you are making anything longer than a single clip. Seven of the ten models accept character reference images, so the same person can appear in shot one and shot six. The other three take a starting frame and improvise from there, which is fine for one clip and falls apart across a sequence.
쇼케이스
Gallery
Script pages lift off a desk and turn into drifting film frames
A ribbon of glowing frames spirals off a rain-slick rooftop
A single written line ignites into a strip of moving images
자주 묻는 질문
자주 묻는 질문
관련